Can confirm that I've seen the same issues 100% reproducible on 4.03+flambda with oasis 0.4.7. Everything worked fine with `< oasis.0.4.6`. I created an issue[1] on the tracker and I also CC'ed the author. It would be nice if you, guys, will also upload your examples to the issue tracker, so that we can fix it ASAP (mine example is too big its an _oasis file that is more than thousand lines of code, I've tried on a smaller example, but bug didn't manifest itself). [1]: https://forge.ocamlcore.org/tracker/index.php?func=detail&aid=1674&group_id=54&atid=291 On Sun, Sep 25, 2016 at 10:20 PM, Boutillier, Pierre < Pierre_Boutillier@hms.harvard.edu> wrote: > DISCLAIMER: I don't know anything about oasis. > > Le 22 sept. 2016 à 17:37, Yann Hamdaoui a > écrit : > > > > I run oasis version 0.4.7 with compiler switch 4.03.0+flambda > It seems indeed that oasis.0.4.7 hates 4.03.0+flambda > Here is my story: > I tried an `opam update` on a debian jessie machine. It had the > consequence to attempt an update of `tls`. (`tls` forces an `oasis setup` > during its ./configure) it fails with `none: unknown option '--enable-lwt'.` > Strange... > So I tried to install tls on a MacOS "just to see" (using switch > 4.03.0+flambda too) There, the compilation of `nocrypto` fails at `oasis > setup` with `E: Unknown flag 'modernity'` By random copy/paste I experience > that the only way to remove this error is indeed to make `Flag modernity` > the last Flag section of the file but in this case other flags seems to be > ignored (let's stay factual: --enable-lwt and --disable-xen are > unrecognized options) > > Then I remembered about your message and I realized that we were both > using +flambda. So I `opam switch 4.03.0 && opam install tls` ... Success! > Finally in my +flambda switch I downgrade oasis to 0.4.6 (on debian > because on MacOS this operation was too much for opam internal solver: it > suggests something implying `===== ∗ 4 ↻ 10 ↘ 15 ⊘ 18 =====` :-)) > Success, I can install tls again! > > I stopped here. I was too happy to be back in a working setting but there > is something to investigate... > > All the best, > Pierre B. > > > -- > Caml-list mailing list. Subscription management and archives: > https://sympa.inria.fr/sympa/arc/caml-list > Beginner's list: http://groups.yahoo.com/group/ocaml_beginners > Bug reports: http://caml.inria.fr/bin/caml-bugs >